The magnitude of the breast cancer problem today emphasizes the recurring and continuing need to determine the best possible predictors of prognosis so that appropriate therapy can be selected on an individual basis for women suffering from this common and dreaded disease. In the United States, breast cancer is the most common malignant neoplasm occurring in women, accounting for about 28 percent of the total cases. There is a noted increase in the number of breast cancers over recent decades. For example, in 1965, it was projected that 1 out of every 17 women (about 6 percent) would develop breast cancer. However, in 1989, the projection for breast cancer development is 1 in 10 women (about 10 percent). This is believed due chiefly to the fact that there are more women and that women are living longer into the cancer-prone years. It will be historically noted that for almost 100 years, the nineteenth century "Halstedian" hypothesis on breast tumor biology was the basis for breast cancer therapy. This hypothesis was a mechanistic concept postulating that cancer spread in an orderly fashion, progressing in a centrifugal manner from the breast to the regional lymph nodes, and ultimately to more distant anatomical sites. It postulated that breast cancers were local at the time of onset; that regional nodes serves as a barrier to local spread; that removal of the involved regional nodes influenced the survival of the patient; and that the type of local-regional therapy employed affected the survival of the patient. Implicit in this hypothesis was the concept that cancer could be held up for a time in the lymph nodes and that adequate removal of the breast with en-block dissection to the regional lymph nodes could effect a "cure." This theory supported the use of the radical and extended radical mastectomies, sometimes with adjuvant radiotherapy, as the best means for obtaining local-regional control. The role of direct hematogenous spread was not considered important; and thus systemic dissemination was not included in the concept [Halsted, W. S., Johns Hopkins Hosp. Rep. 2: 227 (1890-1891)].
The "Halstedian" hypothesis was challenged by MacDonald in 1951 [Sure. Gynecol. Obstet. 92: 443 (1951); Am. J. Surg. 111: 435 (1966)]. MacDonald suggested that human breast cancer was biologically determined from its onset. He stated that he had failed to find any evidence that early diagnosis, size of the tumor, or type of surgery had an influence on the outcome in mammary carcinoma. This report was followed by a period of marked pessimism regarding this form of cancer.
In the late 1970s, an alternative hypothesis to the "Halstedian" concept, based on animal experiments and the failure of some clinical trials to demonstrate the superiority of en-block dissection was advanced by Fisher and his collaborators [Fisher et al., Cancer 46: 1009-1025 (1980)]. The Fisher hypothesis stressed the importance of hematogenous dissemination of the cancer; and postulated that breast cancers were systemic at the time of onset, that regional nodes were not effective as barriers to systemic spread but rather were only indicators of biological behavior and metastatic potential as determined by a complex tumor-host interaction. Moreover, he believed that the removal of involved nodes had little influence on patient survival; and the type of local-regional therapy did not substantially affect survival. His hypothesis suggests that the only impact that could be made on breast cancer was by effective systemic therapy. It supported the use of lesser surgical procedures, with or without radiation, the use of adjuvant systemic chemoendocrine therapy, and the performance of axillary node dissections for staging purposes only.
Present authorities believe that neither of these 20th century hypotheses can be accepted in its entirety. Both modern hypotheses have elements of truth and are involved in explaining clinical events. Thus, these hypotheses should not be considered as mutually exclusive. Some cancers are observed to be very aggressive, producing early disseminated metastases and require adjuvant systemic therapy in addition to the local-regional removal of the tumor. Other cancers are much less aggressive, rarely making a transition to metastatic disease; and can be treated adequately by local-regional therapy without systemic adjuvant therapy. A more complete and detailed review and discussion of these hypotheses and clinical features can be found in The Breast on Comprehensive Management Of Benign And Malignant Diseases (Bland and Copeland, editors), W. B. Saunders Company, 1991, the text of which is expressly incorporated by reference herein.
There are a number of predictive parameters employed clinically for determining the prognosis of patients with breast cancers. Such predictors of prognosis have a variety of different origins and are often used in different combinations to provide a better evaluation and result. One source of predictors is based on the anatomical extent of the cancer in the patient. The predictors thus include staging, tumor size, tumor margins, axillary node status, and tumor location within the breast. A second source of predictors is based on the tumor growth potential (aggressive or virulence). These predictors include invasive quality of the tumor, multi-centricity, histological types, histological grading, growth rate (cell kinetics), the presence or absence of steroid hormone receptors such as estrogen and progesterone receptors, as well as specific biological markers such as carcinoembryonic antigen (CEA) measurements, ferritin, C-reactive protein, acid glycoprotein, alkaline phospatase, silayl transference and urinary hydroxyproline-creatinine ratios.
More recently, the discovery of tumor-specific genes that lead to tumor metastasis has become important in the development of strategies for treatment of breast cancer patients. The existence of specific genes responsible for suppression of tumor metastasis has been reported [Ramshaw et al., International Journal of Cancer 32: 471-478 (1983)]. Several genetic markers are now used in assessing tumor changes and are linked to poor prognosis for the patient. These include amplification of erbB2/HER2/neu gene [Santes et al., Cancer Res. 52: 1916-1920 (1992); and Slamon et al., Science 244: 707-712 (1989)]; and a measured decrease in activity of the nm23 gene [Barnes et al., American Journal of Pathology 139: 245-250 (1991)].
In addition, over the past several years, a number of proteins have been implicated in the aberrant growth of human breast cancer cells. These included: epidermal growth factor receptor [Slamon et al., Science 244: 707-712 (1989)]; p53 protein [Weinberg, R. A., Science 254: 1138-1146 (1991)] a transcriptional factor with tumor suppresser properties; and the nm23 protein, a putative metastatic suppresser [Barnes et al., Am. J. Pathol. 139: 245-250 (1991)].
Today, even better predictors of prognosis are constantly being sought. In 1980, the NIH re-emphasized the importance of reliable prognostic factors by establishing grants to stimulate studies that search for parameters based on histological, histochemical, immunochemical, or other methods that would permit a more precise prediction of prognosis for patients with breast cancers. The ultimate goal is to be able to select the individual patients who can be treated adequately by breast conservation procedures, without any systemic adjuvant therapy; and to define the best type of treatment, both local-regional and systemic, for all other patients so as to offer them the best possible results.
In this continuing search for reliable predictors, there is and remains an urgent need for identification of node-negative patients whose primary tumors have a metastatic potential. The presently known tumor markers, even including evaluation of estrogen and progesterone receptor proteins, do not provide sufficient accuracy and precision and do not offer means for predicting with completely reliability which node-negative breast cancer patients will be likely to have metastatic dissemination. The development of a reliable metastatic marker or indicator, especially in the context of an in-vitro test methodology, would be recognized as a major advance and achievement in this field.